RNA Pseudoknot Project

Literature:

JACS-2011-10094.pdf
JACS-2011-10094-SI.pdf
ParkRNAPseudoKnotVIrtualScreenBioorgMedChem2008.pdf
StelzerRNAvirtualScreeningEnsemblesNatChemBiol2011.pdf



Screening to do list:


Christina's research on RNA pseudoknot (Revised 082712)

  1. repeat Michael's analysis of 1F27.pdb validation using Biotin that has been prepared in Maestro Lig Prep (get Biotin from PubChem not PDB)
    • dock in ICM, GOLD, VINA
  2. do docking of Biotin + Chembridge Diversity in ICM vs. GOLD vs. VINA
  • using structure0001.pdb (unfixed) vs. fixed
    • (do Molprobity on these two structures to compare their Clashscores, etc.)
  • using unfixed 'Best' of the 1,000 structures from Molprobity assignment spreadsheet vs. fixed 'Best'
    • (do Molprobity on these two structures to compare their Clashscores, etc.)




EXTRA: Compare regular docking result of Biotin to those that have also been re-ranked with Schrodinger's Prime software in the Maestro Suite
-- this software calculated the Free Energy of Binding of a ligand:protein complex (or ligand:RNA)
Hypothesis: ???
Testing to see what Prime does to the ranking list. Does it changes significantly?
Ideally we want experimental conditions (Ki, IC50, etc) to compare the ranking list to. (see below)





Christina's research on RNA screening validation (OLD version)
-- don't use our Pseudoknot structures, but rather use known PDB poses (or get ensembles form Molecular Dynamice or NMR data of Stelzer paper)
-- take compounds from Stelzer paper (see above) with known Ki values. Then compare VINA docking score against these.(I thin this is just RNA not a 'psudoknot' exactly - but still useful)
-- can we get the virtual ligands in the supplementary info? Can we get all of the Ki values?
Hypothesis: Virtual Screening with GOLD, ICM and VINA will NOT be able to rank the ligands with a high correlation (Pearson's linear correlation R > 0.6) to experimental data

Compare regular docking results to those that have also been re-ranked with Schrodinger's Prime software in the Maestro Suite
-- this software calculated the Free Energy of Binding of a ligand:protein complex (or ligand:RNA)
Hypothesis: Prime will NOT be able to re-rank the ligands with a higher correlation to experimental data

LATER ????: add in molecular dynamics (using Maestro) of ICM and GOLD docked compounds to further refine selections



Michael's work on Pseudoknot (Spring 12)
For each run, get a Score (either ICM or GOLD) and an image of the ligand docked

COMPARISONS to make
  • ICM vs. GOLD docking of Biotin into 1F27 (this is our POSITIVE CONTROL) - which program does better?
  • Docking of Biotin vs. a 'random' ligand - use either GOLD or ICM results - Is the docking specific?
  • Original vs. 'fixed' pseudoknot docking with biotin in ICM - Did our 'fix' help out?
  • Original vs. 'fixed' pseudoknot docking with biotin in GOLD - Did our 'fix' help out?


ICM- use 'Interactive Docking' mode only - with single SDF ligands as object in the table. Do NOT use the scripts method to run these - fails to properly dock.

  • remove ligand by 'moving object' and then use ligand to define docking site
  1. Biotin into 1F27 (get from PubChem)
  2. Random Chembridge ligand into 1F27

  • (LOWER PRIORITY): - Can ICM find the 'proper' docking site?
  • ICM finds docking site
  1. Biotin into 1F27

  • ICM find binding site by itself
  1. Biotin into our original pseudoknot struct-15
  2. Biotin into our 'fixed' pseudoknot struct-15


GOLD
  • Use sdf files that have been 'converted to 3D' in ICM - after conversion need to go back and manually re-add names to the top line of each ligand since ICM removes them and places an 'm'
  • Define active site as 10 Angstroms around the X, Y, Z coordinates of the Carbon in biotin that is between the 2 nitrogens and the one oxygen (most proximal side of the ligand)
  1. Biotin into 1F27
  2. Random Chembridge ligand into 1F27

  • Define active site as 10 Angstroms around the X, Y, Z coordinates an atom in the site that ICM found
GOLD ACtive site point form ICM (10.004, -1.153, -24.478) NItrogen (C6 N1) Cytosine 6 Nitrogen 1)

  1. Biotin into our original pseudoknot struct-15
  2. Biotin into our 'fixed' pseudoknot struct-15




IMAGES:
1F27 with original Biotin xray pose vs. Biotin from ICM docking
1F27 with original Biotin xray pose vs. Biotin from GOLD docking

TABLE:
Software
Ligand
Receptor
Score
ICM
Biotin
1F27.pdb
-78.02

Random
1F27.pdb


Biotin
Orig PseudoKnot


Biotin
Fixed PseudoKnot
-66.19
GOLD
Biotin
1F27.pdb
55.14

Random
1F27.pdb


Biotin
Orig PseudoKnot


Biotin
Fixed PseudoKnot


Carry out Molecular Dynamics simulations of RNA pseudoknot structures to create an emsemble
Carry out Molecular Dynamics simulations and re-scoring of best ligand poses to re-rank them.(see Malmstrom & Watowich Dengue Virus paper)



Pseudoknot Fix Protocol

STEPS TO FIX STRUCTURE:
take original struct-000X.pdb
Open in PyMol, 'A' tab, Remove Hydrogens
Save Molecule as
struct-000XnoH.pdb
Add Crystal Line
CRYST1 90.000 90.000 90.000 90.00 90.00 90.00 C 1 2 1
(if fail - copy CRYST line from 1F27.pdb and use it)
Save Molecule as
struct-000XnoHcryst.pdb
?? need this step or will Phenix Eng Min without Cryst line??
Copy to a folder on the vdsclass.dyndns.org (vdsclass login, password is same as GDocs)
Go do VDSclass.dyndns.org and open Phenix
(Go to the folder you want Then type in command line 'phenix')
go to Refinement >> ReadySet
struct-0001NoHCryst.pdb -input
Adds H's
-unchekc 'Generate Ligand restraints'
struct-0001NoHcrystReadySet.pdb -output
Select 'Other Tools'
Browse for your file from Ready Set
Choose Energy Minimization
Supply an output file name:
struct-000XnoHcrystReadySetEngMin.pdb

then take this to ICM or GOLD or VINA for docking on the DDFE







Pseudoknot Molprobity Assignment.docx

VDS_6th_Molprobity Assignment Presentation_MichaelEaston_Sp12.pdf

Aims:
  • find binders to an RNA pseudoknot for anti-viral potential

Instructions for Using Molprobity to Analyze Structure:
  • Find Structures here: https://docs.google.com/open?id=0B_Gl3lMyhDsoYTEyNzJhMWEtMDA2Ny00Y2U3LTgzM2ItZWE1MTlmNjM4MmUz
  • Find assigned structures and download
  • Go to : http://molprobity.biochem.duke.edu/
  • On the Browse button - go find your .pdb file.
  • Hit ' Upload'
  • You will see a short page of some results and hopefully see an image on the right hand side (rhs), hit 'Continue'
  • Analyze all-atom contacts and geometry
  • Use defaults that are already selected --> 'Run programs to perform these analyses'
  • Input Data in to the Gdocs spreadsheet under the folder Vdsbiooproject
  • Go back to the Molprobity and Click 'Multi-criterion chart'
  • To save a copy of the page, in your web browser go to File >> Save As >> Web Page complete
  • give it a name that corresponds to the structure you have analyzed. For example: structure-0016.htm
  • Save this file to your desktop and then upload it to the Gdocs folder VDSbiooproject at the bottom as shown
  • You will need to open a new Molprobity for each structure





















FOR STAFF - disregard below:


02/16/12 - received 1000 pdb structures. Uploaded to Google Docs
Michael to generate spreedsheet on GDocs that has a line for each structure and then accompanying Analysis Data from MolProbity.


Fields to save to spreadsheet:
  • Researcher Name
  • Date
  • Name of Structure file e.g. structure-0016.pdb
All-Atom
Contacts
  • Clashscore, all atoms:
  • percentile

Nucleic Acid
Geometry
  • Probably wrong sugar puckers:
  • Bad backbone conformations#:
  • Residues with bad bonds:
  • Residues with bad angles:

Links to Multi-criterion plots for each structure:

Instructions:
click on 'Edit' to edit the page - you must be logged in.
put your cursor where you want to insert the link
go to 'File' button at the top of this wiki page, upload your file and select the 'Click to Link to' button. Then click on your file name.

structure-0016test.htm